The drive to swim is not obvious except as a subgoal to one of the other goals. The the drive to escape is obvious extension/subgoal of the drive to survive.
The drives to cooperate and seek help are not obvious extensions to a single one of the listed drives.
Further, Omohundro’s paper quite explicitly referred to its expectation of sociopathic behavior barring outside influences. It was not that these drives were judged not sufficiently “basic”, it was obvious that they were overlooked.
Cooperation and seeking help will appear in sufficiently advanced AI systems of any design—and to succeed, they both require socially acceptable social behavior.
Cooperation seems unlikely to appear in “sufficiently advanced AI systems of any design”—since to cooperate you need to have some colleagues—and some forms of machine intelligence could well be peerless.
I don’t think Omohundro expects machine sociopaths. He often ends his talks on the topic with a big Buddha slide, like this:
Hopefully, through a combination of understanding our own values and where they came from, together with an intelligent analysis of the properties of this technology, we can blend them together to make technology with wisdom, in which everyone can be happy and together create a peaceful utopia.
The drive to swim is not obvious except as a subgoal to one of the other goals. The the drive to escape is obvious extension/subgoal of the drive to survive.
The drives to cooperate and seek help are not obvious extensions to a single one of the listed drives.
Further, Omohundro’s paper quite explicitly referred to its expectation of sociopathic behavior barring outside influences. It was not that these drives were judged not sufficiently “basic”, it was obvious that they were overlooked.
Cooperation and seeking help will appear in sufficiently advanced AI systems of any design—and to succeed, they both require socially acceptable social behavior.
Cooperation seems unlikely to appear in “sufficiently advanced AI systems of any design”—since to cooperate you need to have some colleagues—and some forms of machine intelligence could well be peerless.
I don’t think Omohundro expects machine sociopaths. He often ends his talks on the topic with a big Buddha slide, like this:
http://www.acceleratingfuture.com/people-blog/?p=2102